The Pea SAD Short-Chain Dehydrogenase/Reductase: Quinone Reduction, Tissue Distribution, and Heterologous Expression

The pea ( Pisum sativum ) tetrameric short-chain alcohol dehydrogenase-like protein (SAD) family consists of at least three highly similar members (SAD-A, -B, and -C). According to mRNA data, environmental stimuli induce SAD expression. The aim of this study was to characterize the SAD proteins by examining their catalytic function, distribution in pea, and induction in different tissues. In enzyme activity assays using a range of potential substrates, the SAD-C enzyme was shown to reduce one- or two-ring-membered quinones lacking long hydrophobic hydrocarbon tails. Immunological assays using a speciﬁc antiserum against the protein demonstrated that different tissues and cell types contain small amounts of SAD protein that was predominantly located within epidermal or subepidermal cells and around vascular tissue. Particularly high local concen- trations were observed in the protoderm of the seed cotyledonary axis. Two bow-shaped rows of cells in the ovary and the placental surface facing the ovule also exhibited considerable SAD staining. Ultraviolet-B irradiation led to increased staining in epidermal and subepidermal cells of leaves and stems. The different localization patterns of SAD suggest functions both in development and in responses to environmental stimuli. Finally, the pea SAD-C promoter was shown to confer heterologous wound-induced expression in Arabidopsis ( Arabidopsis thaliana ), which conﬁrmed that the inducibility of its expression is regulated at the transcriptional level.

The expression of the small pea (Pisum sativum) gene family encoding short-chain alcohol dehydrogenaselike proteins (SAD-A, -B, and -C) is increased dramat-ically by low-intensity UV-B radiation (280-315 nm) and by several abiotic stresses . Compared with other UV-B-regulated pea genes Sävenstrand et al., 2002), the SAD genes are induced more rapidly and at significantly lower UV-B levels. Induction of SAD mRNA accumulation occurs differently in separate tissues . The exposure of pea plants to UV-B results in SAD mRNA accumulation in both epidermal and mesophyll cells of leaves (Kalbin et al., 2001). UV-B treatment also leads to an increase in SAD protein levels in both leaf and stem tissue but not in roots (Scherbak et al., 2009).
SAD forms a tetramer in solution, and this is the sole oligomeric structure found under nondenaturing conditions (Scherbak et al., 2009). The pea SAD proteins are extremely similar: SAD-A and SAD-C differ by only three out of 268 amino acids . SAD-B is more distantly related to the other two isoforms (94% identity to SAD-A; , with the most important difference being the loss of part (12 amino acids) of the predicted nucleotidebinding site. Unlike the coding regions, the promoters of the SAD-A and SAD-C genes differ markedly (Gittins et al., 2002;Supplemental Fig. S1). However, both contain an 11-bp GC-rich motif that was shown to bind nuclear factors in an electrophoretic mobility shift assay (Gittins et al., 2002). This SAD promoterbinding motif (SPBM) has features in common with a number of previously recognized classes of plant ciselements involved in stress-responsive regulation of gene expression, which may explain the induction of the SAD genes by a wide variety of environmental stimuli, including ozone fumigation, wounding, aluminum exposure, and salt treatment .
Sequence database searches using the amino acid sequences encoded by the pea SAD cDNAs revealed homologies to the short-chain dehydrogenase/reductase (SDR) superfamily. The highest similarities were seen with sequences from other plants (Scherbak et al., 2009): (2)-isopiperitenol dehydrogenase from Mentha species (similarity of 55%; Ringer et al., 2005), D 5 -3bhydroxysteroid dehydrogenase (HSD; EC 1.1.1.145) from Digitalis lanata (49%; Finsterbusch et al., 1999), and secoisolariciresinol dehydrogenase from Podophyllum peltatum (46%; Xia et al., 2001). The threedimensional structure of the Podophyllum protein has been solved (Youn et al., 2005), while the Mentha protein is the polypeptide with the highest similarity to the pea SADs for which an enzymatic function is known. The model plant Arabidopsis (Arabidopsis thaliana) contains at least eight different genes homologous to the pea SAD genes, with similarity scores of the translated sequences of approximately 50%. A role in abscisic acid (ABA) biosynthesis has been identified for one of these Arabidopsis SAD homologs, corresponding to gene At1g52340 (enzyme AtABA2; González-Guzmán et al., 2002). For two of the other Arabidopsis SAD gene homologs (At2g47140 and At2g47130), names have been proposed for the corresponding proteins (AtSDR3 and AtSDR4, respectively; Choi et al., 2008) without knowledge of their function.
The SDR superfamily is one of the largest known protein families and comprises thousands of members found in species ranging from bacteria to humans (Oppermann et al., 2003). A large number of SDR x-ray crystal structures have been solved, and typically, these enzymes are either homodimeric (17b-HSD; Ghosh et al., 1995) or homotetrameric (3a,20b-HSD; Ghosh et al., 1994). The proteins are NAD(H)-or NADP(H)-dependent enzymes with a wide variety of substrates, including alcohols, steroids, sugars, and xenobiotics (Jö rnvall et al., 1981;Kallberg et al., 2002).
The physiological substrates of the SAD proteins have remained elusive. In a pilot study, the purified recombinant SAD-A and SAD-C enzymes were incubated with a large number of potential substrates and were shown to catalyze the reduction of carbonyls and quinones at a low rate (M. Brosché and Å . Strid, unpublished data). This suggested that the SAD proteins are catalysts in the metabolism of quinones and/ or carbonyls in response to environmental stimuli. In order to test this hypothesis, the catalytic function of the SAD-C enzyme has now been examined further using different compounds containing these func-tional groups. In addition, to gain further insight into the role of these proteins, their tissue distribution in pea and the expression of pea SAD promoter/reporter constructs in Arabidopsis have also been examined in nonexposed or UV-B-exposed tissue and in wounded leaves.

Substrate Specificity and Kinetics of the SAD-C Protein
The enzymatic activity assays described in this study were all carried out using SAD-C for the following reasons. First, SAD-B, which lacks part of the nucleotide-binding site, was not active using the substrates tested (data not shown). Second, for unknown reasons, the yield of the Escherichia coli-produced recombinant SAD-C protein was much greater than that of SAD-A. However, for the studies described in Supplemental Table S1, which were carried before the SAD-C gene had been identified and cloned, recombinant SAD-A was used.
Substrates for the pea SAD enzyme were sought by measuring the reduction of NAD + /NADP + or the oxidation of NADH/NADPH, respectively, by following the change in A 340 upon incubation with a variety of putative carbonyl-or hydroxyl-containing substrates (Supplemental Table S1), including gibberellins, brassinosteroids, other steroids, alcohols, sugars, flavonoids, quinones, and aromatic carbonyls. Activity was found only with quinones and aromatic carbonyls. p-Quinone reducing activity was detected with NADH as the cofactor using compounds with both one-and two-membered rings ( Fig. 1A) but not those with three-membered rings (Fig. 1B). Permitted ring substitutions included methyl, methoxy, and isopropyl groups. Therefore, compounds such as 2,5-and 2,6dimethylbenzoquinone, 2,6-dimethoxybenzoquinone, duroquinone (DQ), thymoquinone, 1,4-naphthoquinone (1,4-NQ), and menadione (K 3 ) were all reduced ( Fig. 1A). In addition, the aromatic carbonyl 4-nitrobenzaldehyde was reduced. Other quinones with long isoprenoid side chains, such as menaquinone, phylloquinone, ubiquinone, and decylplastoquinone, were not reduced ( Fig. 1B; Supplemental Table  S1). Furthermore, the three-membered ring compound anthraquinone failed to function as a substrate for this enzyme. Other carbonyl-containing chemical species, such as the furocoumarin 8-methoxypsoralen or the monoterpene pulegone, were also not substrates for this enzyme. As expected, the SAD-C protein, NADH, and substrate were all necessary for activity (data not shown).
NADPH could be used as the reductant instead of NADH but with a much reduced reaction rate (data not shown). Also, catalysis of the reverse reaction could be demonstrated (reduction of NAD + by duroquinol and reduced K 3 ) but with lower activity, presumably due to limitation by kinetic or structural factors.
Assuming Michaelis-Menten kinetics, which was supported by typical Michaelis-Menten and Hanes plots (data not shown), catalytic constants were determined for SAD-C protein using three of the compounds that were reduced by this protein: DQ, 1,4-NQ, and K 3 ( Table I). The catalytic rate for these substrates was similar, differing only by up to 15% from an average of 85 nmol NADH oxidized min 21 mg 21 protein. However, the affinity of the enzyme for the three substrates differed markedly, with K 3 having a 3and 9-fold lower K m than 1,4-NQ and DQ, respectively, which was reflected in the k cat /K m (Table I). A twomembered ring (K 3 and 1,4-NQ) seemed to facilitate binding compared with a one-membered ring (DQ), as did methyl substituents on the ring that carries the carbonyl oxygens (compare K 3 and 1,4-NQ). Once bound to the active site, catalysis appeared to proceed at a similar rate for all three substrates. Furthermore, the K m for NADH was determined to 7 mM (data not shown).
It also became evident during the course of this study that all steps in the preparation of the recombinant SAD-C had to be performed in one unbroken sequence in order to obtain the fully active enzyme. Leaving a partially purified protein on ice overnight led to uncharacterized events at the molecular level that severely inhibited function.

Detection of SAD Protein in Pea Tissues by Immunoblot Analysis and Verification of the Protein Identity
Rabbit SAD-A antiserum (which interacts with both SAD-A and SAD-C; Scherbak et al., 2009) was used in immunoblotting to determine the presence of SAD Figure 1. A, Structures of quinones and aromatic carbonyls that function as substrates for recombinant SAD-C protein in the presence of NADH as cofactor. B, Structures of quinones and cyclic and aromatic carbonyl compounds that do not function as substrates for the recombinant SAD-C protein in the presence of NADH as cofactor. DMBQ, Dimethylbenzoquinone; DMoBQ, dimethoxybenzoquinone.
proteins in extracts of different pea tissues. This analysis showed a distribution of the SAD protein in leaves, stems, tendrils, roots, and seeds ( Fig. 2A). Doublets or multiple bands seen on the western blots at the size of the SAD monomer (28-29 kD) represent either tissue differences in the expression of different SAD genes or, more likely, variable posttranslational modifications. Interestingly, dry and soaked pea seeds contained the largest amount of SAD protein, considerably more than any other tissue ( Fig. 2A), which was in accordance with previous results showing that seeds contained the highest levels of SAD mRNA . After soaking, pea seeds were separated into three parts: the embryonic axis, cotyledon, and seed coat. The SAD protein was abundant in the two former parts (Fig. 2B), while the seed coat contained only trace amounts, if any. Significant amounts of SAD protein were also evident in tissue from pods, fresh seeds, and flowers (Fig. 2C). Since both fresh seeds and dry and soaked seeds showed distinct bands on the western blots, it is clear that SAD content is not affected by seed storage. In flowers, the protein was evenly distributed in the wing and standard petal tissue and in the composite keel petal, pistil, and stamen (data not shown). SAD was found in the tissues of all three cultivars tested (cv Greenfeast, Greenshaft, and Kelvedon Wonder; data not shown).
In order to confirm the specificity of the SAD-A antiserum, immunoprecipitation was performed using homogenates of seed cotyledonary axes. Immunoprecipitated protein was separated by SDS-PAGE, and a band of the correct size was excised (Fig. 3A). This was the only band of approximately the right size visible on the gel, demonstrating the specificity of the antibodies. In-gel trypsin digestion and matrix-assisted laser-desorption ionization time of flight mass spectrometry (MALDI-TOF MS) analysis of this band not only confirmed that the labeled protein was a member of the SAD protein family but also showed that it was the SAD-A isoform (  In order to identify which cell types in the examined tissues contained the SAD protein, immunohistochemistry (IHC) using the rabbit anti-SAD-A antiserum was employed to stain formaldehyde-fixed pea tissues.
A preliminary experiment using preimmune serum resulted in no immunostaining (Fig. 4A), confirming that the polyclonal antiserum contained no antibodies cross-reacting with other pea proteins. This preimmune serum was used as a negative control throughout the IHC experiments.
IHC analysis showed some staining of SAD protein in control pea leaf tissues (i.e. those not exposed to UV-B radiation or those exposed to UV-A radiation for 2 or 5 d). The distribution pattern was indistinct, but the protein appeared to be located in the mesophyll and bundle sheath cells around the xylem and phloem (data not shown). Leaves exposed to UV-B radiation, on the other hand, showed more distinct immunostaining concentrated in the palisade parenchyma cells at the upper surface of the leaf ( Fig. 4B; i.e. close to the surface where UV-B radiation impinges on the leaf and therefore has its highest intensity). In order to confirm that the protein labeled in UV-B-exposed leaves was indeed a member of the SAD protein family, immunoprecipitation, SDS-PAGE of precipitated protein, in-gel trypsin digestion of a protein band of the right size, and MALDI-TOF MS analysis were performed. Although it was not possible to discern whether the UV-B-induced protein was SAD-A or SAD-C, this analysis confirmed that SAD was the UV-B-regulated protein detected by the antiserum (Fig. 3C;Supplemental Fig. S2,B and D). As before, no proteins other than SAD were identified by MS analysis of the immunoprecipitate pulled down by the SAD-specific antiserum.
In control stem tissue, minor staining of SAD protein was distributed in the cortex and around the vascular tissue (data not shown). After UV-B irradiation, the SAD protein accumulated in the outer cell layers of the stem, as is clearly seen in Figure 4C. Distinct staining was found in the epidermis or the cell layers below the epidermis and also in the cortex.
The root system, unlike the stems and leaves, showed no noticeable difference in SAD protein distribution between the control tissue and tissue harvested from plants exposed to UV-B (Scherbak et al., 2009). In both cases, there was some minor immunostaining in the root cortex (data not shown). Coiling of pea tendrils is a specific effect of UV-B radiation (Brosché and Strid, 2000). However, SAD IHC analysis of tendril sections showed no clear difference between treatments (controls, UV-A-exposed, or UV-B-exposed plants; data not shown) with respect to the abundance of this protein. Since immunoblotting demonstrated that non-UVexposed pea pods, embryonic axes, and seed cotyledons contained considerable amounts of SAD protein (see above), only the tissues from plants grown under control conditions were analyzed with IHC. In pod tissue, SAD protein was detected around the vasculature (Fig. 4D) and in the thick-walled epidermis of the exocarp (Fig. 4E). Immunostaining was also observed in the parenchymatous cells of the mesocarp and in palisade parenchyma cells. The inner epidermal layer of the pod, on the surface facing the seeds, also contained SAD (Fig. 4F).
The seed embryonic axis (separated from the rest of the seed) is composed of three different tissues: the protoderm, ground meristem, and procambium. SAD protein immunostaining was extremely strong in the protoderm in seeds soaked for 24 h (Fig. 4G). IHC analysis showed very clear and specific staining located in the outer cell layer of the embryonic axis. These were the most heavily stained cells observed in this study, which implies an important role for SAD protein in this tissue.
In the different types of petals examined by IHC, all showed staining in the epidermal cell layer, as exemplified by the keel tissue shown in Figure 4H. This staining was not observed when preimmune serum T, tendril (10.5 mg of material loaded); DSS, dry seed soaked for 24 h in tap water (2 mg of material loaded); DS, dry seed not soaked (2 mg of material loaded); C, purified recombinant SAD-C protein (0.05 mg of protein loaded). B, Presence of SAD protein in different parts of the pea seed after soaking for 24 h in tap water. C, Purified recombinant SAD-C protein (0.05 mg of protein loaded); G, embryonic axis tissue (5.8 mg of material loaded); COT, cotyledon tissue (5.8 mg of material loaded); SC, seed coat tissue (5.8 mg of material loaded). C, Presence of SAD protein in fresh pea reproductive tissues. P Pods (14 mg of material loaded); FS, fresh pea seeds (14 mg of material loaded); C, purified recombinant SAD-C protein (0.05 mg of protein loaded); F, whole flowers (14 mg of material loaded). Molecular masses of protein standard (M) bands are marked in kD. Asterisks denote the monomeric SAD protein bands (28-29 kD).  . The C-terminal peptide proves that the seed-localized SAD is the SAD-A isoform. C, SAD-C sequence, with the peptide sequences obtained by MALDI-TOF MS of band S from lane UL in A underlined. The three amino acids discriminating SAD-C from SAD-A are shown in boldface. The peptide pattern obtained does not identify which of the SAD isoforms is induced by UV-B in leaves. was used instead of SAD antiserum. In the stamen, SAD was predominantly located in epidermal cells (data not shown), whereas in the stigma, this protein appeared to be more uniformly distributed (data not shown). Two bow-shaped lines of cells containing SAD protein were apparent in the pistil ovary (Fig.  4I). SAD-containing cells were also identified on the placental surface facing the ovules.
Heterologous Wound-Induced Transcription Regulated by the Pea SAD-C Promoter in Arabidopsis The above results indicate that SAD has at least two functions in pea plants. The first requires constitutive expression in certain cell layers of different tissue types, such as epidermal and subepidermal cell layers of the cotyledonary axes, pods, stamens, and petals (Fig. 4). The second involves a diffuse but clear increase in SAD content (Fig. 4;Scherbak et al., 2009) in cell layers close to the upper epidermis of leaves ( Fig.  4B) and the epidermis of stems (Fig. 4C) in plants exposed to UV-B radiation. Besides UV-B, other environmental stimuli, primarily wounding, are known to strongly induce SAD gene expression in pea . To more clearly visualize this inducibility of SAD expression and to confirm that the induction is controlled by the SAD promoter, Arabidopsis plants were transformed with promoter/ reporter constructs comprising the SAD-C 0.7-kb proximal promoter (Gittins et al., 2002) fused to either the GUS (gusA; Fig. 4, J and K) or luciferase (luc; Fig. 5) reporter gene.
As judged by the luciferase activity measured in the leaves of nine independent SAD-CP-luc transgenic lines (Fig. 5), it was clear that the 0.7-kb SAD-C promoter drives wound-induced expression in Arabidopsis. For whole tissue visualization of this expression, rosette leaves of SAD-CP-gusA transgenics were wounded by cutting (Fig. 4J) or pricking with a needle (Fig. 4K). It was evident that wound-induced gusA expression, driven by the pea SAD-C promoter, was confined to the locality of the wound site and the neighboring cells.
In contrast to the strong wound-induced expression, the 0.7-kb pea SAD-C promoter was not activated by UV-B exposure in Arabidopsis (data not shown). This indicates either that the 0.7-kb promoter fragment used in the construct does not contain the necessary UV-B regulatory elements or that pea uses different regulatory elements compared with Arabidopsis to direct UV-B-induced expression.
To examine whether Arabidopsis SAD homologs are induced by stimuli, publicly available microarray data in all cases except in A and B, where the dilution was 1:16,000. Specific antibody binding was identified after treatment with biotinylated horseradish peroxidase-conjugated secondary antibody (diluted 1:1) and avidin peroxidase complex, by reaction with DAB substrate. A, Negative immunostaining control leaf from a pea plant exposed to 5 d of UV-B stained with preimmune serum. B, Leaf tissue from a plant exposed to 5 d of UV-B. Note the stained palisade parenchyma cells (arrowhead). C, Stem tissue from a pea plant exposed to 5 d of UV-B. The arrowhead indicates stained cell layers below the stem surface. D, Pod vascular tissue with surrounding parenchyma cells. E, Tissue from the outer part of the pod epidermis. The arrowhead indicates epidermal cells. F, Tissue from the inner part of the pod epidermis. The arrowhead indicates the surface against the ovary cavity. G, Embryonic axis tissue from pea seeds soaked for 24 h showing the protoderm and the ground meristem. Note the obvious difference between the highly stained protoderm cells (arrowhead) and cells of the ground meristem, which have little or no SAD present. H, Keel petal tissue. The arrowhead indicates the stained epidermal cell layer. I, Ovary and ovule (indicated by the asterisk). The arrowheads indicate the two bow-shaped lines in the ovary in which SAD-containing cells are found in addition to cells on the placental surface facing the ovules. J and K, GUS activity in SAD-C P-gusA transgenic Arabidopsis. Rosette leaves were wounded either by cutting (J) or by pricking with a needle (K), then held for 3 h and frozen before staining for GUS activity with 5-bromo-4-chloro-3indolyl-b-glucuronic acid.
were explored. Raw data files were obtained from several databases (see "Materials and Methods") and robust mulitiarray average normalized. To take into account the sample variation in the microarray experiments, parametric bootstrapping combined with Bayesian hierarchical clustering (Savage et al., 2009;Jaspers et al., 2010) was applied (Fig. 6). Of the Arabidopsis genes with highest similarity to pea SAD-C, At3g29250 and At2g47130 (corresponding to protein AtSDR4; Choi et al., 2008) were found to be highly and weakly induced by UV-B, respectively, and overall, these two genes form the closest partners in cluster analysis. In addition to UV-B, both of these genes were induced by osmotic and salt stress and infection with Pseudomonas. Interestingly, neither of these genes was induced by wounding treatment (data not shown).
Since the pea SAD-C promoter is wound inducible (Fig. 5) whereas Arabidopsis SAD promoters are not, this is further evidence that the promoters of these homologous genes contain different regulatory elements. This hypothesis was tested by searching the promoter sequences for recognized stress-responsive cis-elements (Table II). Consistent with its induction by several pathogens, the At2g47130 promoter contained several WRKY recognition sites, previously shown to be important in pathogen responses (Euglem et al., 2000). The SPBM element in the pea SAD-C promoter, which binds nuclear proteins (Gittins et al., 2002), was not found in the Arabidopsis promoters. The SAD-C promoter, but not the promoters of At3g29250 and At2g47130, contained several ACE elements known to be involved in UV/blue light responses (Hartmann et al., 1998). Overall, the promoters of SAD-C and those of the UV-B-induced At3g29250 and At2g47130 differ considerably in the  nature of the previously characterized promoter elements they contain. To identify any novel promoter elements that may potentially be involved in UV-B responses, the three promoters were analyzed using the Motif sampler program (http://bioinformatics. psb.ugent.be/webtools/plantcare/html/Motif_Sampler. html; Thijs et al., 2002). A novel element, AANCAATT (Table II), was found in all three promoters and may be involved in UV-B responses, although further studies are required to determine its exact role.

Seed-Directed Promoter Elements in the Pea SAD-A and SAD-C Promoters
The results of immunoprecipitation and MS analysis (Fig. 3, A and B) showed that SAD-A is the SAD isozyme expressed in pea seeds. Therefore, the SAD-A promoter was searched for known promoter elements conferring gene expression in seeds (Supplemental Fig. S1a). Interestingly, both an RY repeat element (Chamberland et al., 1992;Reidt et al., 2000) and a TACACAT element (with an allowed one-base mismatch; Josefsson et al., 1987) were found. Unfortunately, the published 0.9-kb sequence of the pea SAD-C promoter was too short (Supplemental Fig.  S1B) to allow extensive comparison with the SAD-A promoter.

DISCUSSION
The aims of this study were to elucidate the catalytic function of the pea SDR SAD-C protein and to examine SAD protein localization and the properties of their gene promoters. Chromatographic and electrophoretic analysis of purified recombinant pea SAD-C protein has shown that this enzyme is a tetramer consisting of a dimer of dimers (Scherbak et al., 2009). Functionally, one-and two-membered quinones with short side chains (Fig. 1) were bound by the enzyme and reduced by NADH, with two-membered compounds being bound with higher affinity (Table I). There were no indications in our study that SAD-C is involved in ABA biosynthesis or metabolism (Supplemental Table  S1) similar to its Arabidopsis homolog AtABA2 (At1g52340), with which it shares 40% identity.
All the compounds that were found to be substrates for SAD-C are nonphysiological; the physiological quinones tested (i.e. those containing long hydrophobic aliphatic side chains) did not function as substrates. Several hundred different benzoquinones and naphthoquinones are found in plants (Thomson, 1957(Thomson, , 1971(Thomson, , 1987(Thomson, , 1997, and some of these might function as SAD-C substrates. Unfortunately, none of these are commercially available, and they have mainly been sought and found in medicinal plants. Very little is known about the presence of benzoquinones and Table II.

Gene UV-B Regulation
Motif Name and (Motif Sequence) Yes No -----naphthoquinones, other than phylloquinone, ubiquinone, or decylplastoquinone, in legumes, such as pea, or Arabidopsis. Therefore, we have so far not been able to exactly pinpoint the true physiological substrate(s) for SAD-C, although we now know which classes of compounds are turned over by the enzyme. Immunoblotting demonstrated that the SAD protein is present to a greater or lesser extent in all pea tissues examined (Fig. 2). The pea seed (apart from the seed coat, which contains only negligible levels of SAD) is the tissue with the largest content of this enzyme. In the seed cotyledonary axis, SAD-A was identified as the isozyme present (Fig. 3B). UV-B irradiation of pea plants results in a clear increase in the content of SAD in leaf and stem tissues but not in roots (Scherbak et al., 2009). This indicates that increased expression of the SAD genes, as a result of UV-B radiation, is limited to the exposed tissues and is a local or semilocal effect rather than a systemic response. The conclusion that SAD is locally expressed after stimuli was supported by the pattern of heterologous expression from the pea SAD-C promoter in Arabidopsis following wounding. Only the wound site and cells in the vicinity of the wound show activity of a reporter gene driven by this promoter (Fig. 4, J and K). It could not be conclusively determined whether SAD-A or SAD-C was the isozyme induced in leaves during UV-B exposure (Fig. 3C).
Immunohistochemical analysis demonstrated that in control tissue (as well as in UV-B-exposed leaves and stems), SAD is predominantly located in epidermal or subepidermal cells (Fig. 4, B, C, and E-H). SAD (in this case SAD-A) was most abundant in the protoderm of the pea seed cotyledonary axis (Fig. 4G), which also demonstrates the involvement of this enzyme in development in addition to a role in the response to environmental stimuli. However, expression around vascular tissue was also apparent (Fig. 4, C, D, and H). Although IHC is not a quantitative method, our results support the previous finding (Scherbak et al., 2009) that UV-B increases the SAD content in leaf and stem tissues and also show that this enzyme is mainly confined to the epidermal and subepidermal cell layers (Fig. 4, B and C) that experienced the highest intensities of UV-B radiation (in accordance with previous findings by Kalbin et al. [2001]).
The apparent SAD distribution within the ovary is particularly noteworthy. Figure 4I shows two bowshaped lines of cells in this organ, indicating a specialized tissue producing SAD. The placental surface facing the ovule also contains cells with considerable SAD staining. This again implies an important role for SAD in development and/or reproduction. Therefore, SAD appears to be important in pea both for development and in response to environmental stimuli.
Another interesting finding is that the pea SAD-C promoter directs wound-induced but not UV-B-induced expression in Arabidopsis. This was further examined by identifying UV-B-inducible Arabidopsis SAD ho-mologs in expression studies using public Affymetrix chip data (Fig. 6). Of the eight Arabidopsis genes most similar to the pea SAD genes (Arabidopsis Genome Initiative codes At1g52340, At2g47120, At2g47130, At2g47140, At3g26770, At3g29250, At3g29260, and At3g51680), two, At3g29250 and At2g47130, are induced by UV-B, biotic stress, and osmotic stress. At2g47140 (corresponding to protein AtSDR3; Choi et al., 2008) and At3g51680 show some regulation in response to biotic and/or salt stress. The other five Arabidopsis SAD homologs display only minor responses to environmental stimuli. These different expression patterns exhibited by Arabidopsis SAD homologs suggests that separate members of the pea SAD gene family may also play different roles in response to stress and/or in development, which will most likely be reflected in the cellular distribution of the corresponding proteins. Alternatively, the fact that pea apparently only contains three to six SAD genes  J.R. Gittins and Å . Strid, unpublished data), compared with at least eight putative Arabidopsis SAD genes, may mean that each pea gene has to perform multiple functions, unlike more specialized roles played by individual Arabidopsis genes.
To test whether the promoter regions of the Arabidopsis SAD homologs that were induced by environmental stimuli and the pea SAD promoters share any cis-regulatory elements, the presence of previously characterized light, stress, or pathogen response elements was investigated (Table II). Unlike the pea SAD-C promoter, the Arabidopsis SAD sequences contain no elements previously shown to be important for UV-B-induced expression of the Arabidopsis chalcone synthase gene (Hartmann et al., 1998). Thus, it is likely that the UV-B induction of At3g29250 and At2g47130 is driven by a novel promoter element, which will require promoter deletion analysis for its identification. Furthermore, the pea SAD-A promoter contains elements previously shown to be involved in directing gene expression in seeds (Supplemental Fig. S1A; Josefsson et al., 1987;Chamberland et al., 1992;Reidt et al., 2000).
Whereas the reported immunological studies do not differentiate between SAD-C and SAD-A (Scherbak et al., 2009), except when immunoprecipitation was followed by MALDI-TOF MS analysis (see above), the data on substrate specificity and enzyme kinetics should be regarded as being valid for SAD-C only, although the difference between SAD-C and SAD-A is only three amino acids, all located toward the C-terminal ends of the proteins . These data should be regarded as being of even less relevance for SAD-B, which also lacks a substantial part of the N-terminal nucleotide-binding site.
In summary, the results of this study show that the pea SAD-C protein is capable of reducing benzoquinones and naphthoquinones in vitro. SAD proteins are induced in vivo after environmental stimuli, both at the mRNA and protein levels. SAD antibodies raised against the recombinant protein permitted the localization of this enzyme in different parts of the pea plant and in various cell types within the examined tissues. SAD expression was studied by immunoblotting, immunoprecipitation, IHC, and heterologous expression using the pea SAD-C promoter in Arabidopsis. These findings indicate that SAD proteins function in both development (seeds, pods, placenta) and in response to environmental stimuli, such as UV-B radiation and wounding.

Plant Material
Pea plants (Pisum sativum 'Greenfeast') were grown in vermiculite in a growth chamber at 22°C with a 16-h-light/8-h-dark cycle (approximately 80 mmol photons m 22 s 21 ). Following leaf appearance, plants were watered every 2 d with Hoagland nutrient solution (Hoagland and Arnon, 1950) until they were used for experimentation. In the experiments where flowers, pods, and fresh seeds were studied, the plants were moved to an open growth room and placed under Philips IP54, HPI-T 400-W lamps at 25°C 6 2°C with a 16-hlight/8-h-dark cycle. Two other pea cultivars (Greenshaft and Kelvedon Wonder) were also used to check for variation between varieties. These plants were grown under the same conditions as Greenfeast.
Seeds of Arabidopsis (Arabidopsis thaliana) ecotype Columbia (Col-0) were sown in sterilized compost (70% moss:30% perlite; Weibulls Så-/Pluggjord; Weibull Trädgård). Synchronous germination was promoted by holding at 4°C in the dark for 2 to 3 d. Before transfer to compost, primary transgenic lines derived from Col-0 were selected by germination on Murashige and Skoog basal medium (Murashige and Skoog, 1962) with 0.75% agar (pH 5.7) containing 20 mg L 21 hygromycin. All Arabidopsis seedlings were propagated in a growth room at 22°C with a 16-h-light/8-h-dark cycle. Water and occasional feeding with Hoagland nutrient solution were given during growth.

Purification of Recombinant SAD Protein
Escherichia coli strain BL21(DE3)pLysS carrying the SAD-C overexpression construct was grown and protein expression was induced as described by Scherbak et al. (2009). Three hours after induction, bacteria were harvested, frozen in liquid nitrogen, and stored at 220°C until required. The cells were broken and the SAD protein was purified as described previously (Scherbak et al., 2009). The protein concentration was determined using a Bradford protein assay according to the manufacturer's instructions (Bio-Rad).

Enzymatic Activity Measurements
The catalytic activity and kinetics of the recombinant pea SAD protein were determined by following the change in A 340 caused by the disappearance or appearance of NADH (depending on whether reduction or oxidation was studied) at room temperature. The final reaction mixture (1 mL) typically contained 50 mM Tris buffer at pH 7.5, 0.4 mM NADH or NAD + , 0.3 mg of purified recombinant SAD-C, and potential substrates at appropriate concentrations. When hydrophobic substrates were used, these were dissolved in 5% dimethylformamide. In certain experiments, NADPH or NADP + was used in place of NADH or NAD + to test cofactor specificity. For determination of the K m for NADH, 0.1 mM menadione was used.

UV-B Exposure
Twenty days after sowing, pea plants were exposed to UV-A or UV-A+B radiation, added to their normal illumination, by using Philips TL40/12UV fluorescent tubes. Control plants were exposed to UV-A (0.64 W m 22 ) obtained from the same tubes by using a mylar film filter with a wavelength cutoff at 315 nm. For the UV-A+B exposures, a cellulose acetate film filter was used with a wavelength cutoff at 292 nm. Plants were exposed to UV radiation for either 2 or 5 d (daily exposure equaled 6 h of UV-B irradiation centered around noon) with UV-B BE,300 = 0.3 W m 22 , where UV-B BE,300 was the biologically effective radiation normalized to 300 nm according to Caldwell (1971) and Green et al. (1974). The irradiation was quantified using an Optronics 754 portable high-accuracy UV-visible spectroradiometer (Optronics Laboratories). Following irradiation, tissue samples were collected, weighed, frozen in liquid N 2 , and held at 280°C until required.

Electrophoresis, Western Blotting, and Staining
Plant tissue (0.2-0.3 g) was ground and mixed with 0.1 mL of 50 mM MES buffer, pH 6.3, and centrifuged for 10 min at 13,000g. The supernatant was collected and placed on ice. To the tissue supernatant obtained from the embryonic axis (which contains large amounts of nucleic acids), a small amount of bovine pancreas RNase I powder (Roche Diagnostics) was added. The seed coat, which was difficult to solubilize by the above method, was sonicated on ice for a few minutes to promote protein release. The protein concentration in each tissue extract was determined using a Bradford protein assay according to the manufacturer's instructions (Bio-Rad) with ovalbumin (Sigma-Aldrich) as the standard. SDS-PAGE of protein from pea leaf extracts (10 mg) was performed using 10% precast gels (Ready-Gel; Bio-Rad) and transferred by electroblotting onto Hybond-C pure membrane (GE Healthcare). Western blotting with anti-SAD-A protein primary antiserum (500-fold dilution) was performed according to a standard method (Qiagen). After reaction with alkaline phosphatase-conjugated goat anti-rabbit IgG secondary antibody, immunostained bands were visualized using chromogenic substrate (nitroblue tetrazolium chloride and 5-bromo-4-chloro-3-indolyl phosphate) according to a standard protocol (Promega). Western blots were repeated at least twice with samples from different 20-d-old plants. Where applicable, SDS gels were silver stained using an automated stainer (GE Healthcare) according to the manufacturer's protocol.

Immunoprecipitation
One hundred microliters of pea leaf or pea seed homogenate was mixed with 50 mL of antiserum and 350 mL of RIPA buffer (10 mM Tris-HCl, pH 7.2, 160 mM NaCl, 1% [v/v] Triton X-100, 1% [w/v] Na-deoxycholate, 0.1% [w/v] SDS, 1 mM EGTA, and 1 mM EDTA) and incubated for 4 h at 4°C with agitation. Then, 20 mL of Protein G+-Sepharose bead suspension (Santa Cruz Biotechnology) was mixed with the sample/antiserum solution and incubated overnight at 4°C with agitation. The mixture was centrifuged for 1 min at 1,000g at 4°C, and the pelleted Protein G+-Sepharose beads were washed three times in 1 mL of phosphate-buffered saline (PBS). After the final wash, the pellet was taken up in 50 mL of PBS and mixed with an equal volume of Laemmli buffer containing 10% 2-mercaptoethanol and heated at 100°C for 10 min. After centrifugation at 1,000g for 1 min, the supernatant was separated on a 10% SDS-polyacrylamide gel as described above and the gel was stained with Coomassie Brilliant Blue R-250. The band corresponding to the precipitated protein was excised from the gel and analyzed by in-gel trypsin digestion and MALDI-TOF MS as described by Scherbak et al. (2009).

IHC
Pea tissue samples were placed in cassettes and immersed in Histofix fixation solution (5% formaldehyde in NaPi buffer, pH 7.2; effective osmolarity 300 mosmol; Histolab) for 6 to 24 h. When more than 6 h of fixation was required, the cassettes were held at 4°C. Fixed sections of roots, leaves, tendrils, stems, embryonic axes, and pods were dehydrated by washing in a graded series of ethanol (70%, 90%, 95%) followed by xylene (30%, 10%, 5%), embedded in paraffin at 60°C, and cut into 4-mm-thick sections using a HM 355 S Rotary Microtome (Microm International). The sections were floated on distilled water at room temperature, heat stretched in distilled water at 50°C, and then mounted on positively charged slides (Superfrost plus; Menzel-Gläser). Before immunostaining, the sections were heated to 60°C for 30 min and cooled to room temperature. Subsequently, the paraffin was removed by treatment with xylene, followed by rehydration using a graded ethanol series (95%, 90%, 75%), and finally the sections were immersed in PBS.
The prepared slides were placed in TE buffer (10 mM Tris-HCl, pH 9, and 1 mM EDTA), which was boiled by heating in a microwave oven at 650 W for 30 min. The slides were then cooled to room temperature and placed in 0.1 M PBS. The tissue was blocked by incubation for 1 h at room temperature in a blocking buffer (0.13 Tris-buffered saline, pH 7.6, 1 g L 21 bovine serum albumin, and 0.1% Triton X-100). Primary anti-SAD-A antiserum, routinely diluted 1:8,000 with DAKO ChemMate antibody diluent (Dako Cytomation Norden), was added to the slides, which were then incubated for 1 h at room temperature. Preimmune rabbit serum was used at similar concentrations as a negative control to determine the degree of nonspecific binding. Following three washes with 0.1 M PBS, horseradish peroxidase-linked biotinylated goat anti-rabbit IgG secondary antibody (Dako), diluted 1:1 in DAKO ChemMate antibody diluent, was added to the slides, which were then incubated for 1 h at room temperature. After being overlaid with avidin peroxidase complex, the slides were incubated for a further 1 h at room temperature. Development was performed using a diaminobenzidine (DAB)/peroxidase detection kit (DAKO ChemMate). DAB plus chromogen was diluted 1:51 in substrate buffer containing hydrogen peroxide and added to the slides, which were incubated at room temperature. After the appearance of black stain (approximately 2 min), the slides were washed in PBS, dehydrated using a graded ethanol series (75%, 90%, 95%), cleared in xylene, and mounted with Pertex (Histolab). Positive immunostaining produced an insoluble brown precipitate in the tissues.
Pea seeds (dry, or soaked in tap water for 24 h) and fresh seeds harvested directly from pods were covered with cryoembedding compound (Histolab) and frozen in isopentane and dry ice. Sections of 6 mm were cut using a cryostat at 220°C and then mounted on a slide and stored at 220°C before processing. The frozen sections were immersed in acetone for 10 s and dried at room temperature for 12 to 24 h. The sections were then blocked and incubated with antisera as described above.
Preparation of SAD Promoter-Reporter Gene Constructs, Transformation of Arabidopsis, and Detection of Wound-Induced SAD Expression SAD gene-specific primers were used to amplify fragments from pea DNA following the PCR-based GenomeWalker method (Clontech, BD Bioscience; Siebert et al., 1995). One amplified fragment of approximately 1 kb contained the 5# noncoding and promoter regions of the SAD-C gene (Gittins et al., 2002). A 0.7-kb subfragment carrying the SAD-C proximal promoter was cloned into binary vectors pPCV814luc and pPCV812 (Koncz et al., 1994) to drive expression of the luc and gusA reporter genes, respectively. These constructs were introduced into Arabidopsis Col-0 by Agrobacterium tumefaciens using the vacuum infiltration procedure (Clough and Bent, 1998).
The effect of wounding was tested on leaves from primary SAD-CP-luc Arabidopsis transgenics. Rosette leaves were cut along the midvein; one half was frozen in liquid N 2 while the other was wounded (by pricking with a needle) and held for 3 h before freezing. Extracts made from these samples were tested using a luciferase assay system (Promega). For the SAD-CP-gusA transgenics, rosette leaves were wounded, either by cutting with a blade or pricking with a needle, then held for 3 h, frozen, and subsequently stained for GUS activity with 5-bromo-4-chloro-3-indolyl-b-glucuronic acid (Jefferson, 1987).
Sequence data relevant to this study can be found in GenBank data libraries under the following accession numbers: AF053638 (SAD-A cDNA), AF053639 (SAD-B cDNA), AF097651 (SAD-C cDNA), AF242183 (SAD-A promoter), and AF242182 (SAD-C promoter).

Supplemental Data
The following materials are available in the online version of this article.
Supplemental Figure S1. Potential SAD promoter elements conferring gene expression in seeds.
Supplemental Figure S2. MALDI-TOF MS analysis chromatograms and analysis reports for the SAD content of immunoprecipitates from pea seeds and leaves of pea plants irradiated with UV-B radiation Supplemental Table S1. Substrates tested in enzymatic activity assays with recombinant SAD-A and SAD-C proteins.